Interactive
PRESS
Computing
REVIEW
The following articles are reprinted solely as items of interest for the independent evaluation by members of the Association
of Computer Users, Inc. The opinions, statements of fact, and/or conclusions expressed herein are not those of the Association.
Peter J. Schuyten
Home Computer
Demand Lags
W HAT has happened to the home computer?
What was once thought to be a billion-dollar,
mass-market item that would revolutionize
our lives seems suddenly to have lost its luster.
Even as such giant companies as Texas Instru¬
ments Inc. introduce new systems for the consumer,
industry analysts, market researchers and even
some executives in the field are taking a second look
at the prospects for these, the smallest of computers.
The central question about the market for these
systems seems to be whether the consumer really
wants to pay $500 to $3,000 to balance the checkbook,
educate the children and turn on the sprinklers. At
this point the answer would seem to be no.
t • t
According to knowledgeable analysts, this is a
market that never was. What today goes by the name
of home computer, they say, is and always has been a
hobby product for computer enthusiasts and not the
average homeowner. And when the needs of these en¬
thusiasts are satisfied, the analysts add, the market
will have been saturated. Most manufacturers ap¬
parently realize this and are moving in other direc¬
tions.
For example, when Lewis F. Komfeld Jr., presi¬
dent of Radio Shack, the leading manufacturer in the
field, recently introduced the company’s latest line
of small computers, the TRS 80 Model II, he stressed
that it was not aimed at the consumer market but
rather for the small-business man. “Don’t call them
home computers,” he begged reporters at the news
conference.
Then too, the market research company. Creative
Strategies International, which has long been consid¬
ered bullish on home computers, recently revised its
market forecast downward for next year, from $140
million to $126 million.
With the exception of Texas Instruments, most of
the major manufacturers in the business, like Apple
Computer Inc. and Commodore International, are
exhibiting their wares this week not at the Consumer
Electronics Show in Chicago, but in New York at the
professionally oriented National Computer Confer¬
ence.
“The distinct trend in the personal computer indus¬
try today is that virtually all manufacturers are
moving expeditiously toward serving the profession¬
al, educational and very-small-business markets,”
said Benjamin M. Rosen, an electronics analyst with
Morgan Stanley & Company,
At present, the industry, only about fouryearsold,
is operating in a three-tier market. At the low end,
where most of the 50-odd companies in the business
participate, machines cost about $500, and, because
of their limited hardware and software, are geared
toward home education and entertainment —- video
games. These systems soon will face stiff competi¬
tion from the makers of programmable calculators,
which will be offering sophisticated calculators that
will improve many of the functions.
In the middle tier are machines that sell for about
$1,000, such as Texas Instruments’ 99/4 and the
Apple II computer. These are designed to handle in¬
come taxes and other family finances as well as to
operate appliances. The trouble here is that without
the addition of $1,000 worth of peripheral equipment,
such as disk memory files, printers and maybe a
telephone interface device, they offer little more in
usable computing power than the low-end machines.
At the high end, there are machines aimed largely
at the small-business man. These can cost as much
as $8,000, depending on what additional hardware
and software is purchased, but function much as
commercial minicomputers do, at less cost.
If in fact the home computer market never materi¬
alizes, the need for the Small-business machine is
here today. The accountant, the two-person law firm,
and the dry cleaner chain, for example, are all poten¬
tial users of a machine like Radio Shack’s TRS 80
Model II, which the company will be selling though
its 7,000 retail stores.
The Radio Shack computer hardware includes an
80-character-per-line display, upper- and lower-case
characters, and a memory system that can be ex¬
panded to two million characters.
But the secret to success in this market really in¬
volves software, which is easy to use. The small-busi¬
ness man, even more than the homeowner, does not
want to become a professional programmer just to
use the system. These machines can do the payroll
and perform such additional office tasks as general
ledger accounting and inventory and mailing list
management.
Just as the maker of the low-end home computer is
facing competition in the form of sophisticated calcu¬
lators, manufacturers of such small-business sys¬
tems are bumping up against—but are not yet in—
territory controlled by the commercial minicom¬
puter maker. It remains to be seen whether compa¬
nies like Radio Shack can provide the kind of mainte¬
nance and software support that has been traditional
in the commercial systems business.
Reprinted with permission from The New York Times, June 7, 1979, ©1979 by The New York Times Company.
Information
processing
COMPUTERS
IBM shocks the plug-compatibles
If a new industry can spring up and in
just five years place more than $1 billion
worth of products in the field, can it
disappear just as fast? If it is in the
computer business, and if the producers
are competing head to head with giant
International Business Machines Corp.,
the answer seems to be a qualified yes.
When IBM in January announced its
incredibly powerful and inexpensive
4300 family—a new central-processor
line to replace the aging low- and
medium-range units of IBM’s System 370
line—it delivered what some observers
believed would be a mortal
wound to the so-called plug-
compatible manufacturers
(pcms). These companies make
computers that operate on Sys¬
tem 370 software and are sold to
IBM customers at lower prices. If
anything, the belief that the
pcms were badly hurt is growing
even stronger now among ana¬
lysts who follow the industry.
“The bloom is off the PCM rose
forever,” claims Frederic G.
Withington, senior analyst at
Arthur D. Little Inc. (adl).
Agrees Michael P. Burwen of
input, a Palo Alto (Calif.) mar¬
ket research firm: “The whole
PCM concept is suspect now.”
Losing money. Clearly, the 9 or 10
companies that make up this
nascent industry already are
hurting. Itel Corp., the leader in
units installed, is losing money in
computers now and has seen its stock
tumble by more than 60% in the last
nine months. Amdahl Corp., which
pioneered the PCM concept and built a
$321 million business on it in five years,
recorded its first quarter-to-quarter
sales decline in two years for the first
quarter of 1979. And that happened even
though Amdahl’s big machines do not
compete directly with the smaller 4300.
At least one company that was plan¬
ning a move into PCM computers has
stopped the project. “We’ve given up,”
says Edward Farris, manager of the
Computer Products Div. at Electronic
Memories & Magnetics Corp. (em&m).
“We’re convinced it’s not a viable
market over the long haul.”
But like Mark Twain, most PCM execu¬
tives are saying that reports of their
death have been greatly exaggerated.
“This is still a good business to be in,”
insists Stephen J. Ippolito, founder and
president of ipl Systems Inc., which
expects to ship in July its 100th IBM-
compatible mainframe computer for
marketing by Control Data Corp. De¬
spite the 4300 announcements, Magnu-
son Systems Corp. has been able to win
$10 million in new financing from such
sophisticated investors as Bessemer Sec¬
urities Corp. and Hambrecht & Quist.
And David N. Martin, manager of the
Computer Systems Group at National
Semiconductor Corp., claims that IBM’s
new aggressiveness “makes our strategy
even more valid,” although it has caused
National to hold up production of its
heavily promoted IBM-compatible model.
Surviving. What is obvious to everyone in
the industry is that IBM has drastically
changed the rules under which the plug-
compatible market evolved. With in¬
creased use of advanced technology, with
new pricing relationships among hard¬
ware, software, and maintenance, and
with design concepts that may prove
hard to emulate, IBM has swept aside
previous assumptions about its vulnera¬
bility. “In the old days,” admits Ippolito,
“profit margins were ridiculously high.
You could have low volume and sloppy
management and still survive.” Now, he
figures, “only the strong competitors
will survive.”
In the short run, IBM itself may
provide the makers of small plug-
compatible computers with a golden
marketing opportunity. After the 4300
was announced, IBM was deluged with so
many orders that it had to extend its
acceptance of “first-day” orders through
Mar. 2. Estimates of the order volume
run to 50,000 units and beyond. “Not
everyone will take delivery,” says Bur-
wen of Input, “but even discounted by
60% it’s a staggering number.”
IBM still has not informed customers
when they can expect delivery of the
4300, but the pcms are hoping that if
shipments are delayed by a year or
more, they can help fill the availability
gap. “I think IBM will be overwhelmed by
the market they’ve created,” says Na¬
tional’s Martin. “If they could deliver off
the shelf in 30 days, that would have a
major impact on our viability. But if
they show lead times of one, two, or
s three years, that’s a major plus for us.”
x Little comfort. At best, however, shipping
delays will provide the pcms with only a
1 temporary market niche. “There’s a
difference between a business and an
opportunity,” says Farris in explaining
why em&m finds little comfort in IBM’s
possible delivery problems. Most of the
pcms will try to use that breathing space
to develop strategies that will give them
an edge before an almost-inevitable
shake-out thins their ranks. Withington
of adl predicts that the pcms will have to
change their very nature to survive.
“They’ll have to play the game on the
service and software side,” he says.
Playing catch-up will be a novelty for
the pcms. Taking their cue from Amdahl,
which used advanced semiconductor
technology to outperform IBM’s biggest
computers, other companies have been
able to exploit chinks in IBM’s technologi¬
cal armor up and down the aging System
370 line. By last year, according to esti¬
mates by International Data Corp., the
PCM installed base had reached 434
machines worth $818 million in the U. S.
And that base will more than double this
year, the market research firm predicts.
But with the 4300, IBM has recaptured
technological leadership, in the view of
many analysts. “The pcms can only react
to Armonk [ibm headquarters],” says C.
Oakley Mertz, research vice-president at
International Data. “There’s no way
they can anticipate now.”
Crucial question. The first problem is to
figure out exactly what matching the
4300 in performance involves. “At this
point,” notes one competitor, “we don’t
even know how much of the performance
load is carried by the hardware and how
much by the software.”
This is a crucial question for National
The leading makers
of plug-compatible computers
Amdahl Control Data Itel
A Cmmimrn mmbm of plyg-compatii>l«
mm&il&m installed
Data: International data Corp.
Information
processing/ CONTINUED
Semiconductor, which must decide how
to proceed with its aborted System 400.
Martin believes that the 4300 will create
a huge market for computers that are
“program-compatible” with IBM, but to
take advantage of it, he must first make
sure the 400 is fully compatible with the
IBM machine. “We have to either cut the
price to get it down to the 4331 [the
smallest announced member of the 4300
family] or enhance the performance to
get it up to the 4341 [the top of the
line],” Martin says. The effort to catch
IBM has helped to delay the 400 by six
months, he acknowledges.
The full impact of IBM’s leap in hard¬
ware technology may not be clear until
the first 4300 is delivered and the pcms
have a chance to take it apart. But it
already is clear that IBM has gained some
important advantages in processing
speeds and manufacturing costs with its
concept of packaging high-density logic
chips in a new multilayer ceramic
module. John C. Lewis, president of
Amdahl, points out that the logic chips
themselves use a slower-speed process
than the devices employed by Amdahl.
But Jared A. Anderson, president of
Two Pi Corp., which makes IBM-compati¬
ble minicomputers with roughly the
performance of the smallest 4300 model,
calls IBM’s packaging approach “spectac¬
ular.” So far, he notes, the merchant
semiconductor industry has been unable
to equal iBM’s technique.
The independents will also have a
problem trying to match IBM’s memory
pricing, which is based on extensive use
of memory chips that can each store
65,536 bits (64K) of computer data.
Martin of National, which is primarily a
semiconductor producer, derides the IBM
64K chip as “big, slow, and a power
hog,” and insists that the semiconductor
industry will be “very competitive” with
IBM in memory. But the industry has its
hands full trying to meet demand for the
smaller 16K memory devices, and vol¬
ume production of the 64K is not
expected until 1980.
Matching the changes. Another concern
of the pcms is IBM’s increased use of
secret, microcoded instructions that are
stored on a disk memory and used to
enhance performance and even to change
the operating characteristics of the 4300.
Just how significant this step will be is a
matter of dispute. “What they did was
relatively mild,” says Ippolito. “They
could have put more in microcode.”
But Charles P. Lecht, president of
Advanced Computer Techniques, sug-
National Semiconductor’s Martin: After
IBM’s 4300 debut, rethinking strategies.
gests that increased use of microcode,
coupled with IBM’s new policy of central¬
ized software service, will make planning
difficult for competitors. Lecht points
out that IBM can restructure the proces¬
sor by simply changing the microcode
with a telephone call. “This could cause
previously compatible devices to be
incompatible,” he says.
Changes in IBM policies are also forc¬
ing the pcms to rethink their service
strategies. For example, IBM is now
handling most of its customers’ software
problems over the telephone. Magnuson
has responded by charging a flat rate for
software and hardware maintenance and
counting on the modular design of its
computers to allow software-trained ser¬
vice representatives to handle both
tasks. But Lecht of Advanced Computer
Techniques believes that “it will be hard
for a PCM to establish the same kind of
support network and provide the same
kind of service as IBM.”
Another big jump. Not every PCM is
directly affected by the new IBM family.
In fact, Amdahl’s Lewis points out that
a proliferation of small mainframes
could actually increase demand for large
computers required to control distrib¬
uted-processing networks. “That’s our
niche,” he says. “We just build the
biggest, fastest machine we know how to
build.” So far, Amdahl has managed to
stay ahead of ibm in price/performance.
“The question is,” notes Peter Labe, first
vice-president of Smith Barney, Harris
Upham & Co., who follows Amdahl
closely, “Can they compete with the next
generation?” Labe looks for another big
jump in price/performance when IBM
announces its widely discussed “H
series” of large computers later this year
or early in 1980. But he expects Amdahl
to match it. “I think those guys are
real,” he says. “They’re here to stay.”
Opportunities. One niche left to the pcms
is supplying raw hardware to original-
equipment manufacturers (OEMs), rather
than bumping heads with IBM in the
end-user market. Jean-Michel Gabet,
computer industry analyst for Gnostic
Concepts Inc. points out that so far IBM
has not pursued the OEM business at all,
and he expects Two Pi and others aiming
mainly at OEM customers to survive. Two
Pi’s deal to supply hardware to National
CSS Inc., a time-sharing company that is
entering the systems business, is still
intact, although National css has had to
reduce prices.
Software companies, in fact, may find
a host of new opportunities in the post-
4300 world. “IBM has been shifting the
price burden on its systems from hard¬
ware to software for at least two years,”
points out Mertz of IDC, “and the 4300 is
an incremental step in that trend.”
While IBM still furnishes free of charge
the basic operating systems required to
run its computers, it has started offering
enhancements that can cost hundreds of
dollars per month. “This could create a
market for compatible programs at
cheaper prices,” says one analyst.
On the 4300, IBM’s software pricing
seems almost as aggressive as its hard¬
ware pricing. Lawrence N. Salveson,
president of Software Pursuits Inc.,
which markets a replacement for IBM
operating systems, calls the price of 4300
software “ridiculously low,” although he
concedes that sales in the thousands
might well justify it. But Salveson
insists that the market for independent
products is still wide open. “We can
certainly match IBM’s performance and
we can compete on features,” he says.
As ibm shifts more of the price burden
from hardware to software, the PCMs
will clearly have to follow. “If the small
PCM wants to survive,” says Withington
of adl, “it has to get into software.” But
even for the biggest pcms, this will prove
no easy task. Smaller pcms, strapped for
resources to keep up just in hardware
technology, may have to join forces with
a software specialist. Sums up William
R. Becklean, an analyst for Bache
Halsey Stuart Shields Inc.: “It will never
become impossible to build plug-compat¬
ible products, but it will become increas¬
ingly difficult. This business isn’t going
to be fun any more.” ■
Reprinted with permission from Business Week , June 18, 1979, ©1979 McGraw-Hill, Inc.,
1221 Avenue of the Americas, N.Y., N.Y. 10020, All rights reserved.
A Look At IBM'S DDP Products
Distributed processing has been steadily gaining
momentum over the last few years. In 1975, there was only
an estimated $50 million worth of processing equipment
being employed in a distributed processing environment.
By yearend 1978, the total had risen to over $1.1 billion —
representing a 180% per annum growth. This early suc¬
cess has been due mainly to the efforts of the minicom¬
puter vendors. Their approach has been selling “tools” to
solve problems: thus, when many users began to feel the
need to off-load some of their work to smaller processors,
the minicomputer suppliers were waiting.
IBM, meanwhile, was notably absent from this arena. It
was not until October of 1978 — years after other sup¬
pliers had been successfully implementing DDP net¬
works, that IBM announced its entry into the market with
the 8100 Information System. By this time DDP had
already become established; having gained wide accep¬
tance in the user community. In fact, the words “distrib¬
uted processing” had not been used by IBM in any public
context prior to the 8100 announcement. The White Plains
Computer Manufacturer is undoubtedly planning to make
up for lost time. IBM has made it quite clear that it intends
to become a majorforce in distributed processing —orto
use their nomenclature — “Cooperative Network Pro¬
cessing”.
With the announcement of the 8100, IBM plunged head
first into the distributed processing market. If the appar¬
ent level of orders is any indication, then the 8100 will have
tremendous success. It is rumored that between 25,000
and 44,000, 8100s are on order worldwide. IBM is offering
two models: the 8130 and the 8140. Both machines can
operate in a standalone environment to serve single de¬
partments/locations, or can connect to a System/370 host.
One of the interesting developments has been the
degree of overlap in the DDP offerings between IBM’s
Data Processing Division and its General Systems Divi¬
sion. Originally, DPD marketed products to the larger
user, while GSD concentrated on the smaller user. As time
has gone by, the potential customer bases for each divi¬
sion have begun to overlap more-and-more. The Data
Processing Division introduced the 8100 Information
System and the 4300 family of processors. In the mean¬
time, the General Systems Division, announced the
System/38, plus the availability of quantity discounts on
the Series/1. With the various products which it now calls
distributed processing equipment, IBM can offer a com¬
prehensive set of alternatives for the user community.
Each of the above products may be used in a distributed
processing environment, yet each will retain its own
specific identity. IBM, of course, intends for the 8100 to
become the industry standard distributed processing pro¬
duct. The other products, however, have primary missions
other than distributed processing. The 4300 series, for in¬
stance, is geared toward the replacement of lower end
System 370 installations.
IBM DDP Products Characteristics
8100 SYSTEM
4300 SYSTEM
8130
8140
System/38
Series/1
4331
4341
Model
A23
A 53
3
5
11
K1
Price
$292,940
$46,980
$70,210
$34,920
$68,760
$248,760
Main Memory
512 KB
512 KB
512 KB
12B KB
512 MB
2 MB
Disk
64 MB
64 MB
64 MB
28 MB
64 MB
37 MB
Diskette
Yes
Yes
Yes
Yes
Yes
Yes
Maximum Memory
512 KB
512 KB
1 MB
256 MB
1 MB
4 MB
Maximum Disk
192 MB
320 MB
387 MB
258 MB
1032 MB 1
18,272 MB 2
3270 Use
Yes
Yes
Yes
No
Yes
Yes
Other CRT
2550/3440
2550/3440
1920
1920
1920
1920
Char.
Char.
Char.
Char.
Char.
Char.
Programming
Assembler
FORTRAN
COBOL
Assembler
FORTRAN
COBOL
RPG III
Assembler
FORTRAN
COBOL
BASIC
COBOL
BASIC
COBOL
1. 16 Spindles 2. 32 Spindles
IDC sees the 4300 being used at regional/divisional
branches or operational levels of large corporations, pro¬
viding “host-type” performance and power, as well as
DDP network capabilities. Interestingly, IDC recently
learned that the DDP label was attached to the 4300 only a
few months before the product was announced. Also of
interest, is the fact that many orders for the 8100 have
actually been canceled as a result of the 4300 announce¬
ment. In certain situations, where users would require
maximum capacity from an 8100, they are now opting for
the higher performance capabilities and latitude for
growth which the 4300 offers. In addition, delivery dates
for the 4300 are promised to be much earlier than for the
8100.
In similar fashion, the System/38 announced last
October, is aimed primarily at System/3 users and
System/34 users who are looking to upgrade. Still, this
product may also be considered by IBM users as a poten¬
tial DDP tool. Although the System/38 cannot communi¬
cate with other System/38s, and can communicate only
with specific IBM hosts, it does have certain traits which
make it suitable for various forms of distributed pro¬
cessing.
The Series/1, announced in November of 1976, was
IBM’s first minicomputer. Although targeted at the tradi¬
tional mini market, the Series/1 has had minimal impact-
due in part to early user disappointments. The Series/1
soon began to find its way into DDP environments, where
users felt that distributed processing required a sophisti¬
cated minicomputer, and where they also wanted to stick
with IBM as a supplier. The fact remains, however, that
IBM still sees the Series/1 as a competitor in the tradi¬
tional minicomputer market, even though it is not adverse
to users who wish to implement it in an alternative form of
DDP.
The preceding research on IBM’s DDP Products was
recently published in a memorandum by IDC’s Informa¬
tion Systems Planning Service.
Reprinted with permission from Market Briefs, June 1979, ® 1979 by International Data Corporation.